Google Cloud launches M4N instances for Oracle workloads

Google Cloud has launched the M4N machine series in Google Compute Engine, aimed at high-memory and I/O-intensive workloads.

The release adds a second product to Google Cloud's network- and block-storage-optimised virtual machine line, targeting applications such as Oracle databases, SAP HANA, SQL Server clusters, electronic health record systems and real-time analytics platforms.

Google Cloud is pitching the new instances to customers that currently add extra virtual CPUs to gain more RAM capacity and storage bandwidth. In database environments where software charges are tied to core counts, that approach can raise licensing bills because companies end up buying more compute than the workload requires.

The M4N range uses 5th Gen Intel Xeon Scalable processors and Google Cloud's Titanium offload architecture. Google Cloud says the machines can provide memory-to-core ratios of up to roughly 26GB per vCPU and memory sizes of up to nearly 6TB.

Storage and networking are central to the launch. According to Google Cloud, M4N instances paired with Hyperdisk Extreme can deliver up to 25 GiB/s of aggregate host storage performance and up to 1 million IOPS, while network throughput can reach up to 400 Gbps for VM-to-VM traffic and up to 200 Gbps for internet egress.

The emphasis is on workloads that combine large memory footprints and sustained data movement with strict latency requirements. Google Cloud lists mission-critical enterprise databases, generative AI and retrieval-augmented generation data layers, healthcare systems, ERP deployments, real-time analytics and electronic design automation among the target uses.

Licensing focus

A key commercial argument behind the launch is Oracle licensing. Core-based charging has long made infrastructure sizing a financial as well as a technical issue, particularly for customers that need large memory pools and fast storage access without a matching increase in compute power.

Google Cloud says the M4N design is intended to let customers match compute more closely to actual workload needs while still accessing higher memory and I/O levels. It estimates this could reduce total cost of ownership for Oracle database deployments by more than 20% compared with similar products from rival hyperscale cloud providers.

The machine types scale from 16 to 224 vCPUs and up to 5,952GB of DDR5 RAM. They are offered in three memory-to-vCPU ratio tiers, with predefined virtual machine shapes designed for different workload profiles.

The launch also extends Google Cloud's memory-optimised portfolio, which already includes the M1, M2, M3, M4 and X4 families. Rather than replacing those products, M4N is positioned as a more specialised option for applications where storage and network constraints have become the limiting factor.

Customer response

Google Cloud cited several partners and customers following early use of the new instances.

"Before M4N, meeting our demanding I/O requirements on Google Cloud often required over-provisioning our compute to achieve the necessary performance density. The new M4N instances solve this by delivering high throughput across the smaller to larger shapes," said Sherri Trojan, Senior Principal Solution Architect at Sabre.

Database platform supplier Tessell also commented on the launch.

"We are delighted to see Google Cloud introduce this next-generation high-performance infrastructure for mission-critical database workloads. The new compute platform demonstrates tremendous potential for enterprise Oracle deployments requiring scalability, resiliency, and performance. We are excited about what this innovation means for customers running Oracle workloads on Google Cloud," said Bala Kuchibhotla, Co-Founder and CEO of Tessell.

Intel, whose processors are used in the instances, said the configuration reflected a joint focus on memory and data movement for large-scale database environments.

"With M4N, Google Cloud continues to push the boundaries of platform co-design. By combining 5th Gen Intel Xeon Scalable processors with Google's custom Titanium offload architecture, M4N delivers the extreme memory capacity, high memory bandwidth, and uncompromising I/O throughput required for the world's most demanding mission-critical data environments," Intel said.

Broader push

The launch highlights an increasingly specific approach to cloud infrastructure, with providers building instance families around distinct workload bottlenecks rather than general-purpose compute expansion alone. For enterprise customers running large databases and data-intensive systems, the issue is often no longer raw processor availability but the balance between memory density, storage throughput and software licensing economics.

Google Cloud says the M4N instances are available in selected regions.